Current differential protection is widely used by power system as the main protection for its good performance. The algorithm of a classical current differential protection is shown below.
                    {                                                                                                                                                                  I                        .                                            L                                        +                                                                  I                        .                                            R                                                                                        >                                  k                  ·                                                                                                                                  I                          .                                                L                                            -                                                                        I                          .                                                R                                                                                                                                                                                                                                                                                              I                        .                                            L                                        +                                                                  I                        .                                            R                                                                                        >                                  I                                      d                    ⁢                                                                                  ⁢                    min                                                                                                          (        1        )            
The classical current differential protection is based on full component currents. In the present invention, full component currents are also called total current. Here, full component current is the original current measurement, which consists of both load current and fault component current. And fault component current is the current which is caused only by fault. Load current is not included in fault component current.
The classical current differential protection mentioned above has some problems such as:                1. Low sensitivity and slow operation speed for heavy load conditions or high resistance faults.        2. Need additional function to block the possible mal-trip for external fault with CT saturation.        3. It's relatively difficult to balance sensitivity, speed and reliability in some sense. If we use low setting/threshold for good sensitivity and speed, we will get relatively bad reliability. And if we use high setting/threshold for good reliability, we will get relatively bad sensitivity and speed.        
To resolve these problems, fault component differential protection which is based on only fault component current was proposed. The algorithm of a typical fault component based differential protection is shown below.
                    {                                                                                                                                          Δ                      ⁢                                                                                          ⁢                                                                        I                          .                                                L                                                              +                                          Δ                      ⁢                                                                                          ⁢                                                                        I                          .                                                R                                                                                                              >                                  0.75                  ×                                                                                                        Δ                        ⁢                                                                                                  ⁢                                                                              I                            .                                                    L                                                                    -                                              Δ                        ⁢                                                                                                  ⁢                                                                              I                            .                                                    R                                                                                                                                                                                                                                                                                              Δ                      ⁢                                                                                          ⁢                                                                        I                          .                                                L                                                              +                                          Δ                      ⁢                                                                                          ⁢                                                                        I                          .                                                R                                                                                                              >                                  I                                      d                    ⁢                                                                                  ⁢                    min                                                                                                          (        2        )            
It has better sensitivity and speed than classical current differential relay for heavy load conditions and high resistance fault conditions generally, as the load currents are removed from the calculation.
But existing fault component differential protection only improves the sensitivity and speed in some sense. It does not resolve the reliability problem of CT saturation and it has not a good solution to get good balance between sensitivity and reliability.
Like classical full component based differential protection, the classical fault component differential protection described above in equation (2) may mal-trip during CT saturation if without additional CT blocking functions.
And in addition, it's difficult to get good balance between sensitivity, speed and reliability too. If they improve the sensitivity and speed for internal fault, they will reduce the reliability of external fault. On the other hand, if they improve the reliability of external fault, they will reduce the sensitivity and speed for internal fault.
To resolve the problem above, a prior application PCT/CN2012/076525 (A Method for Identifying the Fault by Current Differential Protection and a Device Thereof) was filed in June of 2012 and disclosed that the additional phase angle information and fault component are used to improve both the sensitivity and reliability of differential protection simultaneously. Furthermore, the new differential protection was immune to CT saturation inherently without any additional blocking functions.
The typical relay criterion in the prior application is shown below.
                    {                                                                                                                                                                    Δ                        ⁢                                                                                                  ⁢                                                                              I                            .                                                    L                                                                    +                                              Δ                        ⁢                                                                                                  ⁢                                                                              I                            .                                                    R                                                                                                                          -                                      I                                          set                      ⁢                                                                                          ⁢                      1                                                                      ≥                                                      F                    ⁡                                          (                                                                        θ                          L                                                ,                                                  θ                          R                                                                    )                                                        ·                                      (                                                                  max                        ⁡                                                  (                                                                                                                                                  Δ                                ⁢                                                                                                                                  ⁢                                                                                                      I                                    .                                                                    L                                                                                                                                                    ,                                                                                                                        Δ                                ⁢                                                                                                                                  ⁢                                                                                                      I                                    .                                                                    R                                                                                                                                                                            )                                                                    -                                              I                                                  set                          ⁢                                                                                                          ⁢                          2                                                                                      )                                                                                                                                            θ                  L                                =                                  arg                  ⁡                                      (                                          Δ                      ⁢                                                                                          ⁢                                                                        I                          .                                                L                                                              )                                                                                                                                            θ                  R                                =                                  arg                  ⁡                                      (                                          Δ                      ⁢                                                                                          ⁢                                                                        I                          .                                                R                                                              )                                                                                                          (        3        )            
Attribute to the adjusting factor of F(θL, θR) based on phase angle information and fault component, the criterion (3) has adaptive characteristic and has very good sensitivity and reliability.
In some special cases, for example, reclosing or power swing, the calculation of fault component current may be difficult or has relatively bigger calculation error. Thus, another criterion based on full component current was also covered by this prior application as shown below.
                    {                                                                                                                                                                                              I                          .                                                L                                            +                                                                        I                          .                                                R                                                                                                  -                                      I                                          set                      ⁢                                                                                          ⁢                      1                                                                      ≥                                                      F                    ⁡                                          (                                                                        θ                          L                                                ,                                                  θ                          R                                                                    )                                                        ·                                      (                                                                  max                        ⁡                                                  (                                                                                                                                                                                    I                                  .                                                                L                                                                                                                    ,                                                                                                                                                          I                                  .                                                                R                                                                                                                                            )                                                                    -                                              I                                                  set                          ⁢                                                                                                          ⁢                          2                                                                                      )                                                                                                                                            θ                  L                                =                                  arg                  ⁡                                      (                                                                  I                        .                                            L                                        )                                                                                                                                            θ                  R                                =                                  arg                  ⁡                                      (                                                                  I                        .                                            R                                        )                                                                                                          (        4        )            
The criterion in equation (4) is based on full component like classical current differential protection. The criterion in equation (4) will satisfy some of the users who prefer classical full component based protection to fault component based protection. And at the same time, the criterion in equation (4) also has adaptive operate-restraint characteristic because of the adjusting factor of F(θL, θR). It is also immune to CT saturation like criterion in equation (3).
But the criterion in equation (4) has less sensitivity than (3) generally; the main reason is that full component current includes load current, thereby, its sensitivity will be influenced obviously by load current especially for the fault cases with heavy load and high fault resistance.
On the other hand, the prior application only covered the adaptive differential protection which was based on pure fault component currents or pure full component currents. While the adaptive characteristic mentioned in prior application can be also implemented by hybrid fault information of both fault component and full component currents instead of pure fault component or pure full component currents.
By this means, the new protection algorithm based on hybrid fault information has the advantages of both full component and fault component and can achieve good balance between reliability and sensitivity.